Computer environment for the dataframes and models
The overall information and modeling environment is being deployed in the Microsoft Azure cloud computing platform. The intent is to facilitate access and maintenance.
Both the VIC and DHSVM model run primarily on a UNIX operating platform (no GUI) and requires users to be proficient in some programming language. The models themselves also have rather complicated structures and algorithms and require the user to have a deep understanding of the underlying physics in order to configure, parameterize and validate. Since DIF functions to disseminate the models and data to stakeholders of varying expertise levels, the problem faced is one where the models have to (a) be easily accessible (b) easy to troubleshoot.
Cloud computing allows us to eliminate the use of standalone machines and to allow all parties equal access to the models and data on a standard platform. Problems (e.g. debugging, validation of results, etc.) can be easily solved if all programs are stored on a “universal” system. The main advantage of having the models and data stored on the cloud is that the cloud is “omnipresent”, i.e. there is no hurdle to accessibility at all hours. The elimination of standalone machines also reduces the cost of maintenance and further technological upgrades if needed (which can prove to be complicated).
Use of the cloud also offers more flexibility to scale the services and costs to the requirements of the project and the amount of resources used. Ultimately, moving computing resources to the cloud would reduce the problem of storage, allow the responsible counterparts to administer model upgrades and allow expansion of accessibility.
A demo version of how VIC and DHSVM runs in the cloud is available.